Key messages
- The developed model enhances the ability to predict KOA, allowing for early preventive measures that could delay the onset of KOA.
- Incorporating serum biomarkers into the predictive model increases its accuracy, suggesting some role in early KOA detection.
- This is the first model that integrates clinical variables and protein biomarkers, making it the perfect option for its practical application, its potential to reduce costs and support personalized medicine.
Study information
Details of the study
“Prognostic model to predict the incidence of radiographic knee osteoarthritis”1
Methods
The study employed a innovative methodological approach, combining clinical variables (age, sex, BMI, WOMAC pain score) and serum biomarkers (APOA1, APOA4, ZA2G, and A2AP were selected based on prior proteomics studies). The model was developed in two phases:
1. Developement phase
In this phase, 282 Caucasian participants from the Osteoarthritis Initiative (OAI) without radiographic signs of KOA (Kellgren and Lawrence (KL) =0 in both knees) were followed-up for 96 months. The outcome was incident rKOA, considered when rKOA with KL≥2 was diagnosed in one of the knees during the follow-up period. Incident rKOA was diagnosed in 29 (10.3%) subjects.
The model’s performance was assessed using the area under the curve (AUC) metric, and internal validation was performed to adjust for overfitting.
2. Validation phase
A cohort of 100 Caucasian participants from the Prospective Cohort of A Coruña (PROCOAC) were included with the same inclusion and outcome criteria and follow-up period. Incident rKOA was diagnosed in 15 (15%) subjects. The AUC was calculated to evaluate the model’s predictive accuracy in this external cohort.
Finally, a nomogram plot was constructed based on the estimated parameters of the proposed model.
Main results
The researchers first developed a predictive model based on demographic and clinical data. This clinical model achieved a moderate ability for predicting rKOA incidence in the follow-up period with an AUC of 0.702. By incorporating ZA2G, A2AP, and APOA1 serum specific biomarkers to the initial model, the AUC reached a 0.831 value. This marked improvement highlighted the role biomarkers may play in detecting early changes associated with KOA.
A simplified version was tested with two most significant biomarkers (ZA2G and A2AP) alongside the clinical variables. High predictive accuracy was maintained with an AUC of 0.820, while simplifying this model and easing its integration into clinical practice. External validation with an independent cohort from PROCOAC showed no statistically significant differences with the proposed models from the OAI cohort.
Conclusion
This study presents a novel, externally validated prognostic model combining clinical and biomarker data to predict rKOA with high accuracy. The developed nomogram is a valuable tool for early stratification of high-risk populations, potentially transforming the approach to KOA by enabling preventative strategies and personalized interventions.
Bibliography
- Paz-González R, Balboa-Barreiro V, Lourido L, Calamia V, Fernandez-Puente P, Oreiro N, et al. Prognostic model to predict the incidence of radiographic knee osteoarthritis. Ann Rheum Dis. 11 de abril de 2024;83(5):661-8.
Link to the full study
This article is a summary based on the following study. For further information and details, please consult the full study. Please do not hesitate to contact us if you have any comments. https://pubmed.ncbi.nlm.nih.gov/38182405/
